** Description changed:
[Target]
Jammy kernel branch: hwe-6.8-next & Noble kernel
[Impact]
After the revert [0] of upstream commit e4731baaf294 ("cpufreq: amd-
pstate: Fix the inconsistency in max frequency units") [1], pushed in
Ubuntu-6.8.0-48.48 / Ubuntu-hwe-6.8-6.8.0-48.48_22.04.1, incorrect value
of cpuinfo_max_freq has been seen when disabling the cpufreq boost in
passive mode (when amd_pstate=passive is passed to the kernel in the
boot command line). This revert was originally introduced to address a
regression observed in older stable kernels (precisely on 6.6.y, see
[2]). However, this regression is linked to the absence of the linux
upstream commit 5547c0ebfc2e ("cpufreq: amd-pstate: Unify computation of
{max,min,nominal,lowest_nonlinear}_freq") that is not backported on
6.6.y but it is present in both 6.9.y [3] and linux-hwe-6.8 [4].
Thus, the revert should not have been applied to linux-hwe-6.8.
[0]
https://git.launchpad.net/~ubuntu-kernel/ubuntu/+source/linux/+git/noble/commit/?id=d76c1093e3fc
[1]
https://git.launchpad.net/~ubuntu-kernel/ubuntu/+source/linux/+git/noble/commit/?id=342e57b2330b
[2] https://lore.kernel.org/all/18882bfe-4ca5-495c-ace5-b9bcab796...@amd.com/
[3]
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=d10692ac4a4a2
[4]
https://git.launchpad.net/~ubuntu-kernel/ubuntu/+source/linux/+git/noble/commit/?id=7a7da95b1e7c
+ [Fix]
+
+ Revert the revert commit that was erroneously included in noble via a
+ stable update including patches from linux-6.6.y.
+
[Test Case]
(Activating the boost)
for cpu in /sys/devices/system/cpu/cpu*/cpufreq/boost; do echo 1 > $cpu ; done
(Deactivating the boost)
for cpu in /sys/devices/system/cpu/cpu*/cpufreq/boost; do echo 0 > $cpu ; done
$ cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq
- [Expected behavior]
+ Expected behavior:
2151000
- [Actual behavior]
+ Actual behavior:
2151
+
+ [Where problems could occur]
+
+ If regressions were to occur they would manifest as cpu frequencies
+ being reported in incorrect frequency units.
--
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to linux in Ubuntu.
https://bugs.launchpad.net/bugs/2109609
Title:
cpufreq amd-pstate: cpuinfo_max_freq reports incorrect value
Status in linux package in Ubuntu:
Invalid
Status in linux source package in Jammy:
Invalid
Status in linux source package in Noble:
Triaged
Bug description:
[Target]
Jammy kernel branch: hwe-6.8-next & Noble kernel
[Impact]
After the revert [0] of upstream commit e4731baaf294 ("cpufreq: amd-
pstate: Fix the inconsistency in max frequency units") [1], pushed in
Ubuntu-6.8.0-48.48 / Ubuntu-hwe-6.8-6.8.0-48.48_22.04.1, incorrect
value of cpuinfo_max_freq has been seen when disabling the cpufreq
boost in passive mode (when amd_pstate=passive is passed to the kernel
in the boot command line). This revert was originally introduced to
address a regression observed in older stable kernels (precisely on
6.6.y, see [2]). However, this regression is linked to the absence of
the linux upstream commit 5547c0ebfc2e ("cpufreq: amd-pstate: Unify
computation of {max,min,nominal,lowest_nonlinear}_freq") that is not
backported on 6.6.y but it is present in both 6.9.y [3] and linux-
hwe-6.8 [4].
Thus, the revert should not have been applied to linux-hwe-6.8.
[0]
https://git.launchpad.net/~ubuntu-kernel/ubuntu/+source/linux/+git/noble/commit/?id=d76c1093e3fc
[1]
https://git.launchpad.net/~ubuntu-kernel/ubuntu/+source/linux/+git/noble/commit/?id=342e57b2330b
[2] https://lore.kernel.org/all/18882bfe-4ca5-495c-ace5-b9bcab796...@amd.com/
[3]
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=d10692ac4a4a2
[4]
https://git.launchpad.net/~ubuntu-kernel/ubuntu/+source/linux/+git/noble/commit/?id=7a7da95b1e7c
[Fix]
Revert the revert commit that was erroneously included in noble via a
stable update including patches from linux-6.6.y.
[Test Case]
(Activating the boost)
for cpu in /sys/devices/system/cpu/cpu*/cpufreq/boost; do echo 1 > $cpu ; done
(Deactivating the boost)
for cpu in /sys/devices/system/cpu/cpu*/cpufreq/boost; do echo 0 > $cpu ; done
$ cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq
Expected behavior:
2151000
Actual behavior:
2151
[Where problems could occur]
If regressions were to occur they would manifest as cpu frequencies
being reported in incorrect frequency units.
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2109609/+subscriptions
--
Mailing list: https://launchpad.net/~kernel-packages
Post to : kernel-packages@lists.launchpad.net
Unsubscribe : https://launchpad.net/~kernel-packages
More help : https://help.launchpad.net/ListHelp
